Yaw drift question ?

[Jdhill]

I am new to this. Had my P3A for about 2 months. I did a IMU calibration after fw update and now when I yaw while in a hover there is a slight drift, when I first got it, it would yaw on a dime. I have now done another IMU calibration ( using a sheet of glass over motor and level), but same results.
I have searched the forums for solutions and all the threads I came up with were for the p2. In those threads they suggest doing a calibration with it shimmed to one side or the other, making it in unlevel for calibration.
I did not find any threads on the subject about the P3. So I'm wondering if I should try this unlevel imu thing or not ? Any advice would be appreciated. Thanks !!

 

[alokbhargava]

A shot in dark: Try calibrating your sticks.

 

[Jdhill]

Thanks for the reply. I tried calibrating sticks.

A shot in dark: Try calibrating your sticks.

g

 

[Jdhill]

A shot in dark: Try calibrating your sticks.

When calibrating sticks, is there a certain pattern to follow when moving sticks ? Maybe I'm not doing it right ?? That's the only thing I can think of. Also thanks for the reply.
I have noticed you are always trying to help those of us that are less knowledgeable, it's greatly appreciated!!

 

[Mako79]

A shot in dark: Try calibrating your sticks.

The sticks have nothing to do with the yaw drift.

Yaw drift is from a poorly installed IMU or from failed calibrations.
Its possible that the main board or the IMU processor unit is installed wonky. The only way to get a perfect yaw drift is to open her up and make sure the main board is dead flat and the IMU box is dead flat.
Another thing that you may consider is doing IMU calibrations when the Phantom is cold so the IMU calibrations apply. It's known that when the Phantom idle for a long time, the board warms up and causes the IMU calibrations to fail or take longer than normal.

 

[Jdhill]

The sticks have nothing to do with the yaw drift.

Yaw drift is from a poorly installed IMU or from failed calibrations.
Its possible that the main board or the IMU processor unit is installed wonky. The only way to get a perfect yaw drift is to open her up and make sure the main board is dead flat and the IMU box is dead flat.
Another thing that you may consider is doing IMU calibrations when the Phantom is cold so the IMU calibrations apply. It's known that when the Phantom idle for a long time, the board warms up and causes the IMU calibrations to fail or take longer than normal.

Thanks for the reply. I did my imu calibration cold and perfectly level using a sheet of glass over the motors, to get it perfectly level, had the app open and on the imu calibration page so as soon as I turned aircraft on I could start.
Unfortunately I think it must be a little off inside. Since I'm not comfortable opening the quad up, I guess I will have to live with it. It only does it when I yaw in a hover, and just a maybe a foot at worse sometimes it's less, but never perfect like it did when I first got it. It started after I did my first IMU calibration ( which was not before my first flight )

 

[ryantrax]

I've seen yaw drift from a bad compass calibration. Have you tried recalibrating the compass?

 

[Jdhill]

I've seen yaw drift from a bad compass calibration. Have you tried recalibrating the compass?

Yes and I have good compass numbers

 

[ryantrax]

Yes and I have good compass numbers

Does it do it in ATTI mode and GPS?

 

[Jdhill]

Does it do it in ATTI mode and GPS?

I am new too this, and not brave enough to try atti mode yet. I fully intend on learning to fly in atti mode, just haven't done it yet!

 

[ryantrax]

I am new too this, and not brave enough to try atti mode yet. I fully intend on learning to fly in atti mode, just haven't done it yet!

Don't blame you there. I'm also leaning toward a wonky IMU install. If you don't want to open it up maybe a warranty claim is in order?

 

[Jdhill]

Don't blame you there. I'm also leaning toward a wonky IMU install. If you don't want to open it up maybe a warranty claim is in order?

Yes maybe, but after 2 weeks of having it I had too send it back for shell cracks, I don't want to loose it again just yet! Having to much fun!

 

[alokbhargava]

Drift comes due to an offset in the control signal to P3 and that can come from Sticks or incorrect calibrations or constantly drifting VPS or GPS positions.

You can correct sticks and calibrations and can switch off VPS but you have no control on GPS values.

Try to increase yaw gain to max and yaw endpoint to min. See if that affects.

 

[Jdhill]

Drift comes due to an offset in the control signal to P3 and that can come from Sticks or incorrect calibrations or constantly drifting VPS or GPS positions.

You can correct sticks and calibrations and can switch off VPS but you have no control on GPS values.

Try to increase yaw gain to max and yaw endpoint to min. See if that affects.

I will give this a try. What affect on flight will this have ( increasing gain and endpoint min ) ?

 

[alokbhargava]

I will give this a try. What affect on flight will this have ( increasing gain and endpoint min ) ?

Increasing gain with become more responsive and might correct the drifting easily. Endpoint with limit the max values. Try combinations of them and let us know if that helps. You have a very typical issue, we are trying to guess the source of the drift. Hope it is not internally through electronic components.

In a close loop control, we never bother for actual values if the reference is solid and steady and actual values are correct. In this case our reference is sticks and actuals values are thru feedbacks from other sensors. Any thing can go wrong. What best we can do is to calibrate so that the deviations are minimum. We can make it to respond faster to adjust corrections by increasing gains and tightening limits.

 

[Jdhill]

Increasing gain with become more responsive and might correct the drifting easily. Endpoint with limit the max values. Try combinations of them and let us know if that helps. You have a very typical issue, we are trying to guess the source of the drift. Hope it is not internally through electronic components.

In a close loop control, we never bother for actual values if the reference is solid and steady and actual values are correct. In this case our reference is sticks and actuals values are thru feedbacks from other sensors. Any thing can go wrong. What best we can do is to calibrate so that the deviations are minimum. We can make it to respond faster to adjust corrections by increasing gains and tightening limits.

Wow! Thanks for the detailed explanation! As soon as it stops raining I will try this and let you know how it works out. It may be a couple days( I hate rain!). Thanks for the help!

 

[Jdhill]

Increasing gain with become more responsive and might correct the drifting easily. Endpoint with limit the max values. Try combinations of them and let us know if that helps. You have a very typical issue, we are trying to guess the source of the drift. Hope it is not internally through electronic components.

In a close loop control, we never bother for actual values if the reference is solid and steady and actual values are correct. In this case our reference is sticks and actuals values are thru feedbacks from other sensors. Any thing can go wrong. What best we can do is to calibrate so that the deviations are minimum. We can make it to respond faster to adjust corrections by increasing gains and tightening limits.

Well I finally got a chance to try this solution.

I increased the yaw gain, played around with different yaw gain settings, but unfortunately none worked. I still get more of a circle when I yaw in a hover. I think it has to be the imu is off inside. I wish I was brave enough to open it up and check, but I'm not!

 

[RusWolf]

OP can you explain exactly how it drifts? So i can go check mine lol

Or better do a video.


Sent from my iPhone using Tapatalk

 

[Jdhill]

OP can you explain exactly how it drifts? So i can go check mine lol

Or better do a video.


Sent from my iPhone using Tapatalk

It doesn't really drift if it's just in a hover. However if it's in a hover and I yaw like a 360. When I first got it and did this it would turn as if there was a pole running through the middle, no movement at all except a spin. Now when I do it, it yaws in about a 1 foot circle.
Hope that explains it. Next time I fly, I will take a video.

 

[Mako79]

Well I finally got a chance to try this solution.

I increased the yaw gain, played around with different yaw gain settings, but unfortunately none worked. I still get more of a circle when I yaw in a hover. I think it has to be the imu is off inside. I wish I was brave enough to open it up and check, but I'm not!

Adjusting gains will not affect or correct the yaw drift.
Gains adjustment should be left default unless you understand how it all works. Changing gains will affect the birds flight characteristics and can cause wobbling and instability and even camera shakes to that lead to "jello" in your videos. Think of gains as in guitar amplifier. If you make the gains too high, it will cause distortion. This is the same with gains in the motor.